I am trying to adapt tensorflow’s example UNet (https://www.tensorflow.org/tutorials/images/segmentation) for my purposes. The main difference is this UNet takes 128×128 images and masks and my images are 512×512 and my masks are 100×100.